A woman walks in the snow during a winter nor’easter storm in the Chinatown area of New York. (Brendan McDermid/Reuters)

As many have noted the forecasts weren’t a bust everywhere! That’s true — especially the ones that hedged lower in their totals. In fact, you can take any of the forecasts from any outlet and find a place where they were awesome. But many failed in important, high population areas, like northern New Jersey, NYC, southern Connecticut and eastern Mass.

The point of this blog was to say that we all have local knowledge and instincts and years of forecasting under our belt, and we need to use it more, especially in cases like this when the models are far from consistent. Forecasters need to trust their good instincts and emphasize the uncertainties — had we done that, people in the Northeast would have known better what conditions to expect.


Up to 14 inches — more in some isolated areas — was in the forecast from Philadelphia to Boston on Wednesday. Winter storm warnings said travel would be “difficult to impossible,” and hundreds of schools closed for the day in anticipation of the snow. Likewise, many workers stayed home and small businesses didn’t open for the day.

In Philadelphia and New York, what ended up happening was about half as bad as what was predicted. In southern New England, the forecast failed miserably.

As a meteorologist, my first instinct is to blame the models. They certainly predicted too much precipitation. They were also inconsistent and confusing from run to run. In particular, the European model stumbled badly when it suggested nearly two inches of precipitation (which would translate to 20 inches of snow) was going to fall south of Baltimore in 24 hours. It showed around 1.5 inches in Washington itself, and what ended up falling was about a half-inch.

Honestly, that was the worst forecast I’ve ever seen out of the European model just 24 hours before storm onset. It looked as if the model was generating erroneously high precipitation because of something called “convective feedback” — a problem that can occur when the model is forecasting strong storm development.

But I can’t just blame the models in this case, because this was really a failure of meteorologists — on TV, in the National Weather Service, working for private companies, whatever. It was a human failure.

We should have known to trust our instincts instead of following the models blindly and hoping for the best. It’s late March. The sun is getting very high in the sky, and daytime temperatures are starting to reflect that. The forecast models were jumping around all over the place right up until the storm started. Forecasters were openly talking about how high the uncertainty was and how “out there” the model predictions were, but that wasn’t reflected in the public-facing forecasts.

Even as it became apparent that snowfall totals were going to come in lower than expected, one forecast office said they were going to “sweat it out” since the models were insisting “it will snow pretty good.”

These are critical forecasts in a critical region, not only because of its population but because of its economy. Even if you assume that just a fraction of the workforce stayed home on Wednesday, that’s still tens of billions of GDP dollars lost. We can’t sweat it out and hope for the best.

Here are some misses that probably had a big impact on peoples’ lives and businesses Wednesday. The forecast as of Wednesday morning is in parentheses.

  • Boston — Maybe 1 inch? (6-8 inches)
  • Worcester, Mass. — 2 inches (4-6 inches)
  • Bridgeport, Conn. — 3 inches (8-12 inches)
  • Waterbury, Conn. — 2-4 inches (8-12 inches)
  • Central Park —  8.4 inches (12-18 inches)
  • East Tremont (Bronx) — 8.1 inches (12-18 inches)
  • NYC/LaGuardia Airport — 9.6 inches (12-18 inches)
  • JFK — 8.7 (12-18 inches)
  • Newark Airport — 8.3 inches (15 inches)
  • Philadelphia — 6.6 inches (12 inches)

Correction: The Southern New England forecast totals have been revised to reflect NWS Boston’s 5 a.m. briefing maps.